1

Identifying the link between socio-

economic disadvantage and participation &

attainment in science: An analysis of the

England National Pupil Database (NPD)

Presentation to NPD User Group, University of

Bristol, 7 September 2016

Professor Steve Strand

University of Oxford, Department of Education

steve.strand@education.ox.ac.uk

01865 611071

2

SES and achievement

• Substantial and long-standing evidence of

association between SES and educational

achievement

– White (1982) Traditional lit review

– Sirin (2005) meta-analysis r=0.31

– True for a wide range of achievement measures,

including Science achievement (Royal Society, 2008)

• Commissioned by EEF & Royal Society to:

– Update evidence re. SES & science achievement

– Support an EEF project funding round

– Report to be published November (delayed)

3

The current study: Three components

• Analysis of NPD to:

– Compare the size of the EVER6 achievement gap from

Foundation Stage (age 5) to the end of KS5 (age 18/19)

using the most recent 2015 cross-sectional data;

– Evaluate EVER6 gaps in progress (both in science and

overall achievement) within each Key Stage;

– Compare EVER6 gap in science to the size of the gap in

other subjects and overall achievement;

– For KS4 chart trends in the size of the EVER6 Science gap

over time; breakdown science achievement by other student

and school characteristics, e.g. by gender, ethnicity, school

type and region.

• Plus two Literature Review elements (causes &

mediators + promising pedagogies/interventions)

4

SES measure used

• EVER6 – Entitled FSM any time last 6 years

• Has limitations, but used because:

– Pupil based not area based;

– The indicator for Pupil Premium (2015/16 £935 for

every secondary and £1,320 for every primary pupil);

– The measure used for reporting in the primary and

secondary school performance tables;

– Utilised by the EEF in their evaluations of

interventions and outcomes.

• Identifies 28% of students age 4-16, though

varies by year group

5

Entitled to FSM and EVER6 percentages

Year group

Number

students in

year group

% Entitled in

Jan. Census

(FSM)

% entitled in

last 6 years

(EVER6) Difference

Reception 635,812 19.0% 19.9% 1.0%

Y1 615,919 19.7% 24.0% 4.3%

Y2 596,582 20.0% 26.8% 6.8%

Y3 583,202 19.7% 28.6% 8.9%

Y4 571,297 19.0% 29.4% 10.4%

Y5 554,717 18.8% 30.3% 11.5%

Y6 535,033 18.5% 30.8% 12.3%

Y7 534,959 18.3% 30.8% 12.5%

Y8 545,967 17.7% 29.9% 12.2%

Y9 560,687 16.7% 28.7% 12.0%

Y10 567,412 15.9% 27.7% 11.8%

Y11 571,906 14.9% 26.6% 11.7%

All pupils 6,873,493 18.2% 27.6% 9.4%

Source: Authors analysis of School Census Jan. 2013

6

Science achievement outcomesKey

Stage Age Mean score Threshold

KS1 7 Science TA (0-3) + Aggregate score

across the 4 Science AT's

Level 2+ / 3+

KS2 11 Science TA (0-6) Level 4+ / 5+

KS4 16 BTEC / OCR Applied Science

Highest GCSE science points score 1+ GCSE science A*-G / A*-C

Entered EBacc Science (& Pathway)

Achieved 2+ GCSE sciences (A*-G)

Ebacc science points score (Mean) Achieved EBacc science (2+ A*-C)

KS5 18/19 Average A level points (Any science) Entered any science A level

Average A level points (Ebacc sciences) Entered any Ebacc science A level

Points score for Biology, Chemistry,

Physics, Electronics, Env. Science,

Geology, Psychology, Computing,

Applied Science

Entry for Biology, Chemistry, Physics,

Electronics, Env. Science, Geology,

Psychology, Computing, Applied

Science

Calculated for the 2012 Y11 cohort using three separate bases:

All Y11 students in 2012 in MMSCH 551,333 (Ever6 proportion = 24.4%)

All Y11 students completing Level 3 342,703 (Ever6 proportion = 17.6%)

All Y11 students completing A levels 216,273 (Ever6 proportion = 13.1%)

Science achievement

7

Effect size measures

• Comparisons across many scales

– NC levels, KS2 test marks, GCSE points scores, A level

grades, % L3+ at age 7, % EBacc science etc.

• Solution - Effect size measures

– Continuous outcomes

• Cohen’s D

• Gap expressed in SD units (sign immaterial)

– Threshold Measures

• Odds Ratio (OR)

• EVER6 as the reference group (OR>1 indicates size of

NonFSM advantage over EVER6)

• ‘Effect’ does not imply causality

8

NPD ResultsKey

Stage Age Measure (D/OR) D OR VA Measure (D/OR) D OR VA

YR 5 - - - - EYFSP score / GLD 0.50 2.3 -

KS1 7 Science TA / Level 2+ 0.43 2.6 0.26 KS1 avge. points /

Level 2+ all subjects

0.51 2.6 0.21

KS2 11 Science TA / Level 4+ 0.46 2.7 0.10 KS2 avge. points /

Level 4+ all subjects

0.49 2.3 0.09

KS4 16 EBacc Points Score /

Entered EBacc Science

0.63 3.0 0.26 Best 8 points score /

5+ A*-C incl. EM

0.71 3.1 0.32

KS5 (all

Y11)

19 Entered any 'A' level

science

- 2.9 - 3+ A level or equiv.

(Level 3) passes

- 2.8 -

KS5 (L3

entrants)

19 Entered any 'A' level

science

1.9 - 3+ A level or equiv.

(Level 3) passes

- 2.2 -

KS5 (A

Levels)

19 Avge Sc. points /

Entered any 'A' lev Sc.

0.30 1.3 0.08 Total A level points /

3+ A level passes

0.41 1.8 0.19

Science Overall attainment

9

Conclusions• EVER6 achievement gaps in science are large (typically

around 0.50 SD and OR= 2.5:1)

• Gap grows when pupils followed longitudinally within a

Key Stage (particularly age 5-7 & 11-16)

• Biggest gap at the end of KS4 (EBacc science D= 0.63

SD and OR= 3:1)

• KS5 gaps largely driven by low attainment age 16 & low

continuation in FTE Post-16. However still poorer

progress 16-19 for those EVER6 who enter ‘A’ level.

• Science gaps broadly the same for other subjects &

overall attainment, not a specific science issue, common

drivers to EVER6 gaps in all subjects

10

International PISA comparisons

• England has a relatively large difference between the score

points of the 5th-95th percentile, only 8/64 countries wider distn.

• But directly assessed the SES achievement gap, though large, is

typical of the international avge. for reading (Jerrim, 2012) and

mathematics (DFE, 2014) e.g. ESCS explains 14.6% of maths score

variance for OECD, 12.4% for England.

Year Reading Maths Science

2006 496 495 516

2009 495 493 515

2012 500 495 516

Notes: Red= <OECD average; no colour= not sig.diff. from OECD average; Green= >OECD average. The average for OECD countries varies, in 2012 was 496 for reading; 494 for mathematics and 501 for science. PISA primary focus 2006=science, 2009=reading ; 2012=maths.

• In last three PISA rounds England’s avge. in science (& problem-

solving) consistently and significantly above OECD avge.

11

KS4 Ebacc science score by region & SES

All students WBRI EVER6 Only

Gap reduces from 9 Ebacc points (1.5 grades in each science GCSE) to 2 points (0.3 of a grade)

12

Policy implications• The earliest measure of science achievement, at age 7, is

strongly predictive of later science achievement (e.g. r=0.60 with

age 11 science), and poor pupils fall further behind over time

• But age 7 science is itself strongly predicted by pupil’s Early

Years Foundation Stage Profile (EYFSP) score at age 5 (r=0.57

with KS1 Science and r=0.68 KS1 APS).

• Suggests interventions to close later science achievement gaps

should focus on learning gaps in the earliest years at school

• Low SES fewer home educational resources, less supportive HLE,

poorer access to high quality child care & pre-schools, struggle with

reading & maths to build vocabulary, general knowledge, scientific

concepts etc. (Sylva, 2014; Morgan et al, 2016).

• Supports Policies like: EY Pupil Premium, free nursery places for

disadvantaged 2yr olds, increasing number qualified staff in EY etc.

13

References• DFE (2014). Achievement of 15-Year-Olds in England: PISA 2012 National Report

(DFE-RR-307). London: Department for Education.

• Jerrim, J. (2012). The Socio-Economic Gradient in Teenagers' Reading Skills: How

Does England Compare with Other Countries? Fiscal Studies, 33(2), 159-184.

• Morgan, P., Farkas, G., Hillemeier, M., & Maczuga, S. (2016). Science achievement

gaps begin very early, persist, and are largely explained by modifiable factors.

Educational Researcher, 45(1), 18-35.

• Sirin, S. R. (2005). Socioeconomic status and academic achievement: A meta-

analytic review of research. Review of Educational Research, 75(3), 417-453.

• Strand, S. (2014). School effects and ethnic, gender and socio-economic gaps in

educational achievement at age 11. Oxford Review of Education, 40(2), 223-245.

• Strand, S. (2015). Ethnicity, deprivation and educational achievement at age 16 in

England: Trends over time. London: Department for Education.

• Sylva, K. (2014). The role of families and pre-school in educational disadvantage.

Oxford Review of Education, 40(6), 680-695.

14

End of Presentation